I want to go faster in (SETI) - 2

[Kruse]

And don't worry about that NT stuff. It works fine in Windows 98.

I'd just like to remind everybody that the command-line (NT) version has a resource leak in Win98. So remember to make sure the client restarts at least once for every WU. Otherwise it will slow down your computer a lot - until it finally crashes (mine could run about 5-6 days before it just stopped responding).

[This message has been edited by Kruse (edited 26 June 2000).]

[MegaManX4]

Now, thats interesting. I didn't experienced that leak yet. I have 2 comps crunching in Win98 and everything is fine, the WU time is the same for all WU's processed on each computer. Even if they run 2 weeks without Seti/Windows restart...

Mega

[Brian R.]

Kruse - Is this a problem for the most recent version, or are you talking about a previous version? Maybe the most recent version has this problem licked.

[Assimilator1]

Maggi

SETIQ as far as I know has never trashed a WU & many of our team use it,your thinking of SETI cache.
Also you can set the level of WU's you want yourself ,& you can set it to manually connect.
Oh yeah ,& you get some extra stats from the SETIQ page

------------------
I guess I won't recruit for my team here TA

[MegaManX4]

We don't use SetiQ, it's evil .
Personely I don't trust programs like this. I always have the feeling that this programs are taking CPU time.

Mega

[Kruse]

The leak I was referring to has been in all CLI versions of Seti I have tried (1.x and 2.x). But as I mentioned in an earlier thread I believe the problem is actually in Win98's virtual NT machine.

Anyway, if you have Win98 and a CLI version of Seti it is very easy to check for the leak. Get TaskInfo2000 (I don't have an URL present), click on the Seti task and watch the handles count in the lower right corner. It should increase with 1-2 handles pr. second (does on my machine and all other Win98 machines I have tried/seen with Seti).

Whether that leak slows down your computer will have a lot to do with how many resources it can handle. I have 256 MB of ram, and as I said, after less than a week my computer consistently stopped until I started using a batch file that restarts Seti after each WU (like Maggi's).

[Mark F]

Well that explaines alot to me! SETICache won't go to the next WU on one of my machines (no idea why, and have tried many cominations of settings ; any ideas), unless/untill I move the mouse. So I went back to useing CLI version, that I didn't want to use because it would crash every so oftain (didn't know why, and 'nothing' else everdoes, and not useing SETICache). So I'm stuck between a rock and a hard place. Use CLI and crash every so oftain , or use SETICache and let it sit idol when I'm not around at the end of a WU ; no good choise.

I'll look into SETIQ (or will it send all my WUs to AT ). Or can I use batch files to do the same? I'll be going on holiday soon (and boy do I need it; you know your too bussy when you don't have time to check the forum regularly), so copying MAGGI's is not an option, it must be "fully" automaited; ie able to UPLOAD and DOWNLOAD without help, and deal with SETI servers being down.

Any ideas, help or sugestions?

Sorry for the rambling and so many questions

Thanks,
Mark F.

------------------
OH NO, my retractable cup holder swallowed a DVD...
and burped out a movie




[This message has been edited by Mark F (edited 28 June 2000).]

[Maggi]

Hey Mark,

feel free to use my "24/7-online" version ...

:seti

seti-1.exe -stop_after_process
seti-1.exe -stop_after_xfer

goto seti

Copy'n'paste the above to a batch file and link it to your autostart folder ...

I use two of those simultaneously here at work on my dual CPU machine and it works pretty well.



Maggs

[Mark F]

Thanks MAGGI,

I'll start with that. If I understand/remember my DOS days. It will loop if the the SETI servers are down; try to xfer till it times out, try to run but has no info, try to xfer again, repeat till it works. Is this correct?

So I assume it could be expanded to be used with more cached WUs, so that when a xfer times out it will go to the next WU, then start over. Is this correct?

Also I take it "seti-1.exe" sould point to my SETI executable?

Thanx again,
Mark F.


------------------
OH NO, my retractable cup holder swallowed a DVD...
and burped out a movie

[Maggi]

yup sort of like that ...

I have two subdirectories in which there's the batch and the executable along with my user files.

The loop above will restart the exe twice for each WU, so in case one is finished, it'll stop, the rerun SETI to send it, the rerun SETI to start crunching the new WU.

Strange thing is, that no matter which option you use for SETI, it will either send/receive or crunch a WU depending on whether you have a WU on your disk ..

Example ...
no WU on disk
-stop_after_process option
> will receive a WU and start crunching

WU on disk
-stop_after_xfer option
> will crunch it and after that also send & receive

Weird, hu ?

Anyway, of course you could extend it to the batch I posted above for caching WUs manually.

Cheers,
Maggi

[Assimilator1]

MMX4

It uses 0.04% on my Cel @ 550 MHz measured using Taskinfo 2000.Hardly anything.

Here's the link BTW
http://www.iarsn.com/index.html#/register.html

Here's another 1
http://atm.idic.caos.it/



------------------
I guess I won't recruit for my team here TA

[Rattledagger]

Maggi, this isn't correct, and the help from 2.4.i386-winnt-cmdline.exe isn't correct.

-stop_after_xfer does only transfer result/gets new wu. It does NEVER process a unit.
If it had processed the wu too, your send_seti.bat for uploading/caching wouldn't work.

-stop_after_process processes to end of a wu and stops. If no wu/already done wu, gets a new one and processes this before stopping.

For always connected people, use -stop_after_process if you're not using caching.
For caching, use a loop of render_seti.bat followed by send_seti.bat.
It's no point to use -stop_after_process followed by -stop_after_xfer at once, because if no problem the cache is refilled at once, but if connection down, seti@home "will retry in an hour", effectually blocking execution of next in cache.

[Maggi]

Ok Rattledanger,

that is how it's supposed to be ...

Try the -stop_after_xfer option when there is already a WU in your directory ...
It'll crunch the WU and then send it.

Try the -stop_afetr_process option when there is no WU in your dir ...
It'll download one and starts crunching it.

The reason why I use both option in a row is to make sure that the problem that Kruse mentioned won't affect your rig too heavily, that is to say it'll restart seti twice per WU.

Cheers,
Maggi

[Maggi]

PS: my send_seti.bat works only if there's either a result file or no WU !!!

try it out and you'll see

[jms]

Maggi,
The client I use(2.4) responds correctly to the -stop_after_xfer option, if there is a WU in the directory or not.

Jan